Summary:The consequences of using two pre‐emergence herbicides on autumn‐sown perennial ryegrass (Lolium perenne L.) were measured over 3 years in a field experiment. Ethofumesate (2‐ethoxy‐2,3‐dihydro‐3,3‐dimethylbenzofuran‐5‐yl melhylsul‐phonate) and methabenzthiazuron [N‐(benzothiazol‐2‐yl)‐N, N'‐dimethylurea] were each applied at 0, 0.8, 1.6 and 3.2 kg ha−1one week after drilling the grass seed. Monthly counts of plants per unit area and tillers per plant were made until the following April and the herbage dry weights of the major species were recorded at seven harvests.Heavy infestations ofStellaria media(L.) Vill. andPoa annuaL. occurred on untreated plots. Ethofumesate, and to a lesser extent methabenzthiazuron, proved effective in suppressing these weeds, only low rates being needed to controlS. media. Both weeds needed to be controlled to achieve full benefits to the crop; control resulted in a 4‐fold increase in ryegrass tiller density by the first spring and a 6‐fold increase in ryegrass herbage was recorded at the first harvest in May. Some yield benefits were carried over into the second year but not into the third. In total, an extra 4.16 t ha−1of ryegrass dry‐matter was harvested from plots treated with ethofurnesate at 0.8 kg ha−1compared with the controls. The results are discussed in terms of defining optimum crop densiti